CICD For Data Science & Machine Learning - Whiteboard
CICD for data science is confusing. I'm on the whiteboard to go through a basic overview.
Before the fun starts, you must:
-- Build something cool, i.e., your ML project
-- Check your project into source control, i.e., Google Docs for code
Then we begin.
As the DS/ML, you're not responsible for building your company's CICD system, but you do need to format your projects such that you can use them. This means git-based workflows, automated builds & tests, and separate environments, i.e., dev/test/prod.
In the whiteboard video I step through these concepts and try to break things down.
At Continual, we care a lot about CICD for ML because it's typically in Jenkins, CircleCI, Travis, Harness, and Github-Actions that organizations take source code and turn it into a deployable state where a human (for ml projects) decides whether to deploy it or not. The more we can follow good software engineering practice, the easier and more structured our ML deployments will be.
Смотрите видео CICD For Data Science & Machine Learning - Whiteboard онлайн без регистрации, длительностью часов минут секунд в хорошем качестве. Это видео добавил пользователь Gus Cavanaugh 06 Январь 2023, не забудьте поделиться им ссылкой с друзьями и знакомыми, на нашем сайте его посмотрели 1,066 раз и оно понравилось 20 людям.